In this paper, I presented the method to show similarities between paired individuals by using connected graph and applied it to show the similarities of districts in the central region of Tottori prefecture regarding composition ratios of agricultural departments. In this study, I calculated the dissimilarity index between paired districts by using an average absolute value of differences in the percentages of the number of farms producing products belonging to each of five departments from which they earn most farm receipts, occupied in the total number of farms. In this method, the dissimilarities are expressed based on the shortest distances between paired districts on the connected graph, which is adopted as a result of deletion of edges one after another from the complete graph in such a manner that an edge, which realizes highest correlation coefficient between the dissimilarity index and shortest distance on the connected graph for the paired districts after deletion, is chosen for each deletion stage. As the districts are connected through edges in a connected graph, the similarity relationship of districts can be shown systematically through the manner of connections among districts compared with usual methods applying multivariate analyses. In case of the application to the region, factors influencing the similarity were also reflected on the difference in branches of the connected graph. Therefore, the method is useful to analyze the similarity relationship systematically among individuals such as districts.
